DICOM tags and metadata: what the fields mean
A DICOM tag is a numbered field stored inside the file that describes the image, the patient or the exam. Every tag has a two-part number, like (0010,0010), and a value. Together the tags are the file’s metadata.
This post explains how tags are organised and which common ones are worth recognising. It does not explain what the images show. That is for the patient’s doctor.
How a tag is written
A tag is written as two four-digit hexadecimal numbers in brackets, separated by a comma. The first is the group and the second is the element.
For example, (0010,0010) is group 0010, element 0010. It holds the patient’s name. Groups gather related fields. Group 0010 is mostly about the patient, group 0008 about the exam and the image in general, group 0020 about identifiers and position, and group 0028 about how the pixels are stored.
Each tag also has a name, a value type and a value. Most software shows the name so that you do not need the numbers.
Common tags
These ones come up most often.
| Tag | Name | What it holds |
|---|---|---|
| (0010,0010) | Patient’s Name | The patient’s name |
| (0010,0020) | Patient ID | An identifier assigned by the clinic |
| (0008,0060) | Modality | Type of equipment, such as CT, MR or CR |
| (0020,000D) | Study Instance UID | A unique ID for the whole study |
| (0020,000E) | Series Instance UID | A unique ID for one series |
| (0028,0010) | Rows | Image height in pixels |
| (0028,0011) | Columns | Image width in pixels |
Modality is a short code. CT stands for computed tomography and MR for magnetic resonance. Other codes include CR and DX for X-ray. For how studies, series and images relate, see CT vs MRI, series and study.
Study, series and image
DICOM organises images in layers. A study is one visit or exam. A series is one run of images within the study, such as one set of slices. An image is a single slice or frame.
The Study Instance UID and Series Instance UID are how software knows which files belong together. A UID is a long string of numbers separated by dots. It is not meant to be read by people. It only has to be unique, so that two files with the same UID are known to be part of the same study or series.
The tags that describe pixels
Some tags tell the viewer how to draw the picture. Rows and Columns give the size in pixels. Others give the pixel spacing, the number of bits per pixel and how the grey values are stored. Window values, if present, suggest a starting brightness and contrast. See window and level explained.
You rarely need to read these yourself, but they explain why two files of different sizes can show the same body part, or why an image looks washed out until the window is changed.
Identifying information is in the file
Patient name, ID, birth date and sex are stored as tags inside the file. They are not only in the file name. A file renamed to scan1.dcm still carries them. The same goes for dates, the institution and sometimes the name of the physician or operator.
This matters when you share a study. Anyone who opens the file with a tag viewer can read these fields. See DICOM viewer privacy.
Tags are not always filled in
A tag can be present with an empty value, or missing altogether. Some exports leave out fields, and some replace the real values with placeholders. A missing tag does not mean the file is damaged. Two different scanners may fill in different sets of tags for the same kind of exam.
Do not edit tags in a file you may need to use clinically. Changed metadata can stop a clinic from matching the study to the right patient. If you think a field is wrong, tell the clinic that produced it.
Reading tags in a viewer
Radix DICOM has a metadata panel that lists tags for the open image, grouped as patient, study, series and image. It is a way to look at what the file says about itself, such as the modality or the image size. It does not interpret anything, and it is not a medical device or a diagnostic tool.
Getting the most from the metadata
- Use the modality and series description to tell which set of images you are looking at.
- Check Rows and Columns if an image looks smaller or larger than expected.
- Treat the patient fields as private data, because they are.
- Leave interpretation to the patient’s doctor.
If a field seems wrong or a study appears mixed up with another, do not try to fix it yourself. Contact the imaging centre and ask them to check the export.